In studies of lake-atmosphere interactions, the fluxes of momentum, water vapor and heat (latent and sensible heat) are parameterized as being proportional to the differences in wind, humidity and air temperature between the water surface and a reference height above the surface. The proportionality or transfer coefficients are often assumed to follow the gradient observation above lake surface or the parameterizations established for the marine atmospheric surface layer. Optimization against the eddy covariance and micrometeorology measurements made over a large shallow freshwater lake (Lake Taihu) shows that the transfer coefficients of momentum (C_D10N), water (C_E10N) and heat (C_H10N) were C_D10N=1.52×10^-3, C_E10N=0.82×10^-3 and C_H10N=1.02×10^-3, respectively. These values are in good agreement with the values derived from the eddy covariance measurement in other inland lakes. Comparison with oceanographic parameterizations suggests that lake surfaces were aerodynamically rougher than open oceans under similar wind conditions, which may due to the shallow depth of the lake, and that these parameterizations can bias the annual lake evaporation estimate by as much as 40% higher. Our results also suggest that these coefficients can be regarded as constants independent of stability and wind speed. Sensitivity analysis indicated that the mean error between calculation and observation of latent heat flux (LE) decreased 0.5 W/m², that of sensible heat flux (H) decreased 0.4 W/m², and no difference between friction velocity (u*) calculation and observation if stability correction was considered since around 83% of data were in neutral condition. If the effect of wind was considered, the mean error between calculation and observation decreased 0.004 m/s for u*, increased 1.3 W/m² for LE, and negligible for H. This study can provide reference for the research on lake-atmosphere interaction.